Skip to content
← Tools / Skills
ToolExa/10 trustVerified
Exa Python SDK logo

Exa Python SDK

Official Python client for Exa, the neural search API built for AI.

Visit official site →

About

exa-py is the official Python SDK for Exa (formerly Metaphor), a neural search engine that indexes the web as embeddings instead of keywords, designed specifically for AI agents and RAG. The SDK wraps every endpoint — neural search with multiple types (auto, neural, deep variants), page contents extraction, find-similar, cited LLM answers, streaming search, and async deep-research tasks. Exa is backed by Lightspeed, NVIDIA's NVentures, and Y Combinator ($22M total raised, TechCrunch July 2024) and powers search in tools like Cursor.

What it can do

  • **search()**: neural search with types auto/neural/deep/deep-lite/deep-reasoning, domain filters, date ranges, category (company/people)
  • **contents**: highlights (query-relevant passages), full text, summaries, subpages crawling via get_contents()
  • **answer() / stream_answer()**: LLM answer with citations over search results
  • **output_schema + system_prompt**: structured JSON or text output on any search type; stream_search() yields OpenAI-style chunks
  • **research API**: async long-form research tasks (research.create / poll_until_finished)
  • **MCP server**: exa-labs/exa-mcp-server (5.1k stars) exposes the same API to AI assistants

Setup

```bash

pip install exa-py

```

```python

from exa_py import Exa

exa = Exa() # reads EXA_API_KEY from environment

# or: exa = Exa(api_key="your-api-key")

result = exa.search(

"blog post about artificial intelligence",

type="auto",

contents={"highlights": True},

)

for r in result.results:

print(r.url, r.highlights)

```

Get a key from dashboard.exa.ai/api-keys.

Prerequisites

Python 3.9+ An Exa API key (set EXA_API_KEY or pass api_key=...).

Use cases

RAG retrievalresearch agentsanswer enginescompetitive intelligence

Works with

claudesmolagentslangchaincustom

Trust score — 8/10

/10 trustVerified

Scored 8/10 (verified tier, metrics-v1). Evidence: 234 GitHub stars (log-scaled vs catalog peers); blended with ~5,406,331 monthly package downloads; last push 7d ago; 17 releases/12mo; 49 open issues. Manual review: repo is exa-labs/exa-py (not exa-python); $17M Series A verified via TechCrunch; later rounds excluded as unverified; named customers (Cursor/OpenAI/Anthropic) provider-claimed. License: MIT.

★ 234 GitHub stars

How the trust score is built

Full methodology →
Adoption · stars/downloads, log-scaled vs peers
6.3/10
Maintenance · recency, releases, issue load
8.2/10
Openness · license + docs
10.0/10
Manual review · sanity read + red-flag check
8.0/10
  • · 234 GitHub stars (log-scaled vs catalog peers)
  • · blended with ~5,406,331 monthly package downloads
  • · last push 7d ago; 17 releases/12mo; 49 open issues
  • · manual review: repo is exa-labs/exa-py (not exa-python); $17M Series A verified via TechCrunch; later rounds excluded as unverified; named customers (Cursor/OpenAI/Anthropic) provider-claimed

Sources

Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.